← McKinsey Interview Insights

McKinsey·Software Engineer·Technical Phone Screen·Intermediate

Intermediate
Apr 2026

Summary

Interviewed for a data science role at McKinsey. One question but it was the kind that sounds deceptively basic until you're actually sitting there trying to justify your answer.

Questions Asked (1)

Q1

In what situations would you use the median rather than the mean as a measure of central tendency?

Product Analytics & MetricsTechnical Trade-offs
Author's notes

I started with the skewed distributions angle (income, housing prices) which is the obvious answer, and they nodded along.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Start by defining mean and median and their respective strengths. Then, explain that the median is preferred when data is skewed or has outliers, as it provides a more robust measure of central tendency. Finally, illustrate with a concrete example relevant to software engineering or product analytics.

Pro tip: Mention that the median is often used in A/B testing for metrics like response times or revenue per user, where outliers can distort the mean and lead to incorrect conclusions.

1. Define the measures

Briefly define mean and median, highlighting that mean is sensitive to extreme values while median is resistant.

2. Identify when to use median

Explain situations such as skewed distributions, presence of outliers, or when data is ordinal. Emphasize robustness.

3. Provide examples

Give concrete examples, e.g., household income, website load times, or user session durations, where median is more representative.

4. Relate to role and company

Connect to software engineering and McKinsey context: e.g., analyzing performance metrics, customer data, or A/B test results where outliers are common.

5. Summarize trade-offs

Conclude that while median is robust, mean is useful for further statistical analysis; choice depends on data distribution and goal.

Key Points to Mention

  • Skewed distributions (e.g., income, response times)
  • Presence of outliers that can distort the mean
  • Ordinal data where mean is not meaningful
  • Robustness of median to extreme values
  • Use in A/B testing and performance metrics
  • Trade-offs: mean uses all data, median focuses on central position

AI-generated suggestions, not part of the candidate's original notes. May be inaccurate — verify before relying on them.